{"id":"W7077883051","doi":"10.48448/yapw-z370","title":"SimSMoE: Toward Efficient Training Mixture of Experts via Solving Representational Collapse","year":2025,"lang":"en","type":"other","venue":"Underline Science Inc.","topic":"Geochemistry and Geologic Mapping","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"Artificial Intelligence in Medicine (Canada)","funders":"","keywords":"Scalability; Representation (politics); Similarity (geometry); Routing (electronic design automation); Artificial neural network; FLOPS; Constant (computer programming); Training (meteorology)","routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false,"invisible_to_affiliation_only":false},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaepi_narrow"],"consensus_categories":[],"category_scores_codex":[0.0007149467,0.0002845557,0.0003802313,0.0004950367,0.0002070415,0.0001285647,0.002151172,0.000221302,0.0004135515],"category_scores_gemma":[0.0006036503,0.000260544,0.00009211595,0.001461549,0.0006630817,0.0001130625,0.000796086,0.0002146051,0.0000106885],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00007285883,"about_ca_system_score_gemma":0.001300443,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0001068146,"about_ca_topic_score_gemma":0.0000152217,"domain_scores_codex":[0.9971399,0.00004634053,0.0004142231,0.0009901784,0.0009000967,0.0005092486],"domain_scores_gemma":[0.9981158,0.0001718303,0.000382938,0.0008502466,0.0003407366,0.000138478],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"not_applicable","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.00003483406,0.001013953,0.0005481084,0.001477283,0.0003066296,0.0002897737,0.020047,0.05004354,0.06856466,0.1854288,0.5309545,0.1412909],"study_design_scores_gemma":[0.001077605,0.0001378033,0.0001763732,0.00137681,0.00004711497,0.0000920002,0.002131191,0.666113,0.01769527,0.01272835,0.2969244,0.001500102],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"other","genre_gemma":"other","genre_scores_codex":[0.00006336111,0.0004184201,0.4205495,0.002449559,0.001152901,0.0002963142,0.0000210598,0.0002557447,0.5747932],"genre_scores_gemma":[0.3905247,0.00004657415,0.1887381,0.0008998987,0.0005531975,0.00004719878,0.00005884674,0.00005264437,0.4190789],"genre_candidate":"other","genre_consensus":"other","teacher_disagreement_score":0.6160694,"threshold_uncertainty_score":0.9999847,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.0246899707046746,"score_gpt":0.2786816864818582,"score_spread":0.2539917157771836,"validation_status":"score_only:v0-immature-baseline","note":"Baseline scores from an immature model (maturity gate not passed). Scores rank; they never assert a category."}}